I thought I could feed it an MP4 video file and have it spit out a subtitle file. Nope, apparently it must play the video file through in real time somehow and pick up the sound through the microphone. It also misses a trick, if it really does AI voice recognition and translation, of feeding it a speech MP3 or WAV file and spitting out a TXT file. Since all the functionality should be inside it. Nope. Verdict: nope.